direkt zum Inhalt springen

direkt zum Hauptnavigationsmenü

Sie sind hier

TU Berlin

Inhalt des Dokuments

Interactive Perception


Action is a crucial part of the perceptual process. Humans interact with their environment constantly to reveal information and use the interaction to understand the consequences of their actions in the world. Interactive Perception exploits these ideas to build autonomous mobile manipulators that use interactions to perceive, understand and manipulate their environment. mehr zu: Interactive Perception

Representation Learning


Truly versatile robots cannot rely on representations that are specifically hand-crafted for every task. Instead they must be able to learn these representations from experience. mehr zu: Representation Learning

Manipulation Planning


Brushing teeth, eating, opening a door: humans interact with the environment countless times each day without thinking of it. Manipulation planning requires the understanding of Object, Robot, and Environment (ORE) interactions. We research the "ORE triangle" to advance the field and shift the manipulation planning paradigm toward new and exciting realms. mehr zu: Manipulation Planning

Soft Hands


We design, build, and use soft robotic hands, like the RBO Hand 2. Made from rubber and actuated pneumatically, they are highly compliant and extremely robust. With them we can reliably grasp objects with uncertain geometries and positions, by offloading some perceptual difficulties to the morphological computation of the hand itself. mehr zu: Soft Hands

Motion Generation


How to find motion strategies that are robust in the real world? We combine control and planning under uncertainty to generate motion for mobile manipulators. mehr zu: Motion Generation

Protein Structure Prediction


What shape does a protein take when it folds? This questions have preoccupied researchers for decades. We are building algorithms that leverage experimental and evolutionary data to answer this question, and predict the structures of proteins from their sequences. mehr zu: Protein Structure Prediction

Protein Motion


Understanding the function of a protein requires understanding the protein's motions behind that function. We work towards developing a "computational microscope" that simulates these motions computationally efficient and biologically accurate. mehr zu: Protein Motion

Human Grasping


Can we improve robotic grasping by studying how humans manipulate and grasp objects in their environment? By varying subjects' visual and touch capabilities, we characterize their grasp strategies and transfer them to robots. mehr zu: Human Grasping

Heuristics for Ball Catching


Is it better to use heuristics or optimization for control problems such as ball catching? mehr zu: Heuristics for Ball Catching

Zusatzinformationen / Extras


Schnellnavigation zur Seite über Nummerneingabe

Hier geht es zu Youtube